JP3512698B2 - Voice packet multiplex transmission apparatus and voice packet multiplex transmission method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice packet multiplex transmission apparatus and a voice packet multiplex transmission method for packetizing voice information into code information of a fixed information block and multiplexing and transmitting the first packets of a plurality of channels. .

[0002]

2. Description of the Related Art Conventionally, in the transmission of voice, the system is being transferred from an analog system to a digital system, and in the case of telephone transmission by voice, which requires real-time property, the voice is A / D converted due to digitalization. The packet is divided into a predetermined number of bits, transmitted as a packet through the line, D / A converted, and transmitted to a target destination.

This voice packet transmission device is described in Japanese Patent Application Laid-Open No. 5-304509 as a first conventional technique. In the publication, in the related art, in each line corresponding unit, the voice information digitized by the bit string conversion circuit is packetized as it is by a predetermined length of 256 bytes by the pulse coding (PCM) method, and the packetizing multiplex unit is used. When the total packet amount exceeds the transmission line capacity, some packets on an arbitrary line are discarded. In this discarding, in the voice information digitized by the PCM method, the higher the bit, the larger the weighting of the amplitude of the voice signal, and if the same amount of information is missing, the higher the bit, the lower the clarity of the voice.

Therefore, in the publication, as shown in FIG.
When the total packet amount exceeds the transmission capacity, the packet having the smallest bit position number is preferentially discarded.

FIG. 4 is a block diagram showing an outline of the configuration of the voice packet transmission device of the publication. The voice packet transmission device of FIG. 4 inputs the input voice information digitized by the PCM method for each line, divides it into a predetermined length, and edits the bits in order from the lowest position to the highest position. The bit string conversion circuit 201 and the control unit 206 as a bit string conversion unit for dividing into two parts, the bit position number addition circuit 202 and the control unit 206 for adding bit position numbers to the output of the bit string conversion circuit 201, and the input voice information are input. The silence detection circuit 205 for detecting whether or not the bit section at each position is silent, the control unit 206, and the control information for the sound section in the output of the bit position number addition circuit 202 based on the detection result of the silence detection circuit 205. A packetizing circuit 204, a packet information generating circuit 203, and a control unit 206, which discards the silent section and outputs the packetized packet.

Further, as shown in FIG. 5, the multiplexer 230 detects the bit position number for each line from the output of the packetizing circuit 204, and based on this detection result, when the total packet amount exceeds the transmission line capacity. Includes a buffer 231, which preferentially discards the packet having the smallest bit position number, a bit position number detection circuit 232, a control circuit 234, and a packet multiplexing circuit 233 which multiplexes and outputs the outputs of the bit position number detection circuit 232. Includes and.

The input audio signal 221 digitized by the PCM system is input to the input terminal 211, and the bit string conversion circuit 201 inputs the input audio signal 2 from the input terminal 211.
21 is divided into a predetermined length (256 bytes). The bit string conversion circuit 201 performs bit conversion on this 256-byte audio information based on the bit string conversion control signal 223 from the control circuit 206. Bit position number addition circuit 2
02 adds the bit position signal 224 output from the control circuit 206 to the 32-byte information for each bit position input from the bit string conversion circuit 201, and outputs it to the packetization circuit 204. The packetizing circuit 204 adds control information such as a source address and a destination address created by the packet information generating circuit 203 to the information for each bit position input from the bit position number adding circuit 202 to packetize the information. . On the other hand, the silence detection circuit 20
5 receives the input audio signal 221 and inputs the bit string conversion circuit 2
When the voice information in the section processed in 01 is silent, the packet discard signal 222 is output to the packetization circuit, and the packetization circuit 204 outputs the packet discard signal 222.
If is input, the packet is discarded and the packetized information is transmitted to the multiplexing unit 230.

The multiplexer 230 stores the packet transmitted from the line interface 210 in the buffer 231, and the buffer 231 receives the packet and then receives the packet reception signal 24.
1 is supplied to the control circuit 234. The packet once stored in the buffer 231 is stored in the bit position number detection circuit 232.
The bit position number detection circuit 232 detects the bit position number 242 added by the line corresponding unit 210 and supplies the bit position number 242 to the control circuit 234. The control circuit 234 receives the sum of the packet amounts input from the line corresponding unit 210. If the number exceeds the transmission capacity, one of the lines having the smallest bit position number 242 added by the bit position number adding circuit 202 is selected, and the packet discard signal 243 is sent to the bit position number detecting circuit 232 of that line. Output. If this results in
If the total sum of the packet amounts to be multiplexed, that is, the total packet amounts from the line corresponding unit 210 excluding the discarded packets exceeds the capacity of the transmission path, the packet discard processing is similarly performed for the other lines. Be seen. It should be noted that the packet discarding process should not be continuously performed on the same line as much as possible. In this way, the packet including the bit close to the least significant bit of the PCM original information is preferentially discarded.

This is because, in the code by the PCM system, the closer to the least significant bit, the smaller the amplitude component of the voice, and thus the error amount of the signal component when the packet is discarded is determined. I am less.

Further, as a second conventional technique, Japanese Unexamined Patent Publication No. Hei 6 (1994) -6
In Japanese Patent No. 6383, the configurations of the line-corresponding transmitter and the line-corresponding receiver are the same as those shown in FIG. 4, and when the voice packets of some lines are discarded due to line congestion or the like. , In order to reduce the deterioration of the voice clarity on the receiving side, when the total packet amount exceeds the transmission capacity, the voice type information output from the voice type information detection circuit gives priority to the voice packet whose vowel part. Had been discarded.

[0011]

However, in the voice packet transmission apparatus of the first prior art described above, the voice coded data of the coding method such as the PCM method in which the strength of the signal component does not directly correspond to the bit string is used. When packetizing a packet, it is possible that the loss of voice quality when a packet composed of a lower bit position of data is discarded is less than that of a packet composed of a higher bit position. Not exclusively.

As for the second method of discarding a voice packet of the second prior art, when a voice information is divided into a vowel part and a consonant part and a coding method other than packetization is applied, for example, When a packet is generated by an efficient compression coding method or the like, the priority of packet discard cannot be determined by the same standard. Further, in any of the conventional methods, the probability of packet discard occurring in each multiplexed channel is statistically equal, so that in a situation where packet discard occurs frequently, voice quality of all channels deteriorates. there is a possibility.

An object of the present invention is to discriminate between a voiced part and a silenced part of voice information inputted from a line, and then packetize and transfer only the voiced information and the silence detection information. In addition to making efficient use of the transmission line, when the traffic of voice packets temporarily increases, the packets of the new conversation channel are temporarily stored in the packet buffer and the multiplexing timing is delayed, so that the voice due to packet discard is discarded. When the packet is discarded because the quality deterioration is minimized and the packet buffer capacity has been exceeded, the sender is notified of this, and the continuation of the non-call state and the deterioration of the overall communication quality are minimized. The aim is to reduce the number.

[0014]

A voice packet multiplex transmission apparatus according to the present invention generates a voice packet from voice information input from a line in a line-corresponding transmission section except when in a silent state, and sets it in a limited band of a transmission line. Based on the packet multiplexing control means for controlling the packet multiplexing within the range of the maximum number of multiplexing, the voice packets generated in excess of the maximum number of multiplexing are multiplexed until the traffic of other voice packets decreases. A packet buffer means for temporarily storing voice packets for delaying, and a signal for notifying the receiving side of the line that this packet is discarded when the packet is discarded because the capacity of the packet buffer means is exceeded. And a packet discard notifying means.

Further, according to the present invention, in a voice packet multiplex transmission device for multiplexing and transmitting voice packets via a line, a voice / voice signal for discriminating a voiced portion and a voiceless portion from voice information input from the line is transmitted. The silent part is encoded, and the sound part is encoded based on the discrimination result of the sound / sound detection part, and the silent part generates the silence detection information only at the start of silence, and in the silent state thereafter, the code is generated. A voice coding unit that does not perform coding, information that identifies a channel and information that identifies voiced / unvoiced with respect to the coded voice information output from the voice coding unit and the silence detection information. A packetizing unit for adding and packetizing the packet into a first packet, and the first packet output from the packetizing unit.
The first to delay the start of multiplexing of
Shape and the packet buffer section for storing the packet, the second packet turned into multiplexing the first packet output from the packetizing unit or said packet buffer of
And a second packet assembling unit that outputs the second packet to a transmission line and is multiplexed by the second packet assembling unit.
Monitoring the number of channels corresponding to the first packet,
Delay the start of multiplexing based on the number of channels in
A packet multiplexing control unit for controlling storage of a first packet of a channel to be stored in the packet buffer unit and for controlling the start of multiplexing of the first packet stored in the packet buffer unit. And the first packet
Notification of the discarding of the first packet to the sender of
Push button (PB) signal for responding to the sender
And a PB signal generator for outputting to the line
The packet multiplexing control unit transfers the first packet to the packet
When storing in the buffer section, empty in the packet buffer section.
If the first packet is discarded when there is no available capacity,
Together, the PB is connected to the line corresponding to the sender of the first packet.
The PB signal generator is instructed to output a signal .

Further, the present invention is a voice packet multiplex transmission method for multiplexing voice packets for transmission via a line , wherein a voice part is included in voice information input from the line.
And a silent part are discriminated, and the voiced part is discriminated based on the discrimination result.
The part is encoded and output as voice information,
The silent part is coded only at the start of silence.
Outputs sound detection information, and then encodes in the silent state.
First of all, for the voice information and the silence detection information,
Information that identifies the channel that represents the line and that that identifies voiced / silent
Packetized into the first packet by adding
The start of the multiplexing of the packetized first packet
Stores the first packet for delay and
Packetized first packet or the stored packet
The first packet is multiplexed and cellized into ATM cells for transmission.
It is output to the transmission path, and is paired with the first packet to be multiplexed.
The number of corresponding channels is monitored and the number of channels in the monitoring is monitored.
Based on the number of channels to delay the start of multiplexing.
It controls the accumulation of 1 packet, and is also accumulated.
Controlling the start of multiplexing of the first packet according to
There is no free space to store
If not, the first packet is discarded and the first packet is discarded.
Output the PB signal to the line corresponding to the sender of 1 packet
Characterized in that it.

[Operation] According to the present invention, referring to FIG. 1, the packet multiplexing control means 30 includes a voice packet output from the line corresponding transmission section 10 corresponding to each line, and a packet buffer means. The number of voice packets already stored in each 20 is monitored, and a voice packet newly generated exceeding the maximum multiplexing number or a voice packet of a line already stored in the packet buffer means 20 corresponds to the line. When output from the transmission unit 10, the voice packets are stored in the packet buffer unit 20 as much as possible.

Here, when voice packets exceeding the capacity of the packet buffer means 20 occur, the packet multiplexing control means 30 instructs the line-corresponding transmission section 10 to discard those voice packets. At the same time, it instructs the packet discard notifying means provided in the line reception handling unit 60 to send a signal for notifying the packet discard to the receiving side of the line corresponding to the discarded packet.

Further, when the number of voice packets to be multiplexed falls below the maximum multiplexing number, the packet multiplexing control means 30 multiplexes the voice packets accumulated in the packet buffer means 20 and not yet transmitted. Start.

As described above, when the traffic exceeding the limited bandwidth is temporarily generated, the start of packet multiplexing of the newly generated traffic is delayed, so that the communication quality is deteriorated due to packet discard and delay fluctuation. As much as possible. Also, when discarding a packet, the sender is notified of the packet discard, and the deterioration of the communication quality of other communication is minimized.

[0021]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described in detail with reference to the drawings.

[First Embodiment] (1) Description of Configuration A first embodiment of the present invention will be described with reference to the drawings.

Referring to FIG. 1, in the first embodiment of the present invention, on the sending side to the transmission path 2, the sound / voice of the means for judging the voice / non-voice state of the voice information input from the line is detected.
On the basis of the sound detection section 13 and the sound / sound detection determination result of the sound / sound detection section 13, the sound information is encoded in the sound state and the silence detection information is generated only at the start of the sound in the silence state. Then, the speech coding unit 11 of a unit that does not perform the subsequent coding, the coded speech information of the speech coding unit 11 and the information for identifying the channel and the information for identifying the sound / silence in the silence detection information. A packetizing unit 12 as a means for adding and packetizing is provided in the line corresponding transmitting unit 10 of the line corresponding packet transmission processing unit.

On the other hand, when a packet exceeding the preset multiplex capacity occurs on the transmission line 2, the packet buffer section 20 of means for accumulating the packet in the buffer is delayed in order to delay the timing of multiplexing, and An ATM cell assembling unit 40 as means for multiplexing packets from the line and packets stored in the buffer and sending them to the transmission line.
A decision is made to accumulate packets in the packet buffer section 20 based on the voice / silent state of each line and the multiplexing capacity of the transmission path, and a decision to start multiplexing of untransmitted packets in the packet buffer section 20. Packet multiplexing control means 30 for making a decision to discard a packet from the available capacity of
Have. Here, when outputting to the ATM network, the AT
Although the M cell assembling unit 40 is used, a public line network or I
When using the SDN line network, the packet buffer unit 2
The packet stored in 0 is used as the first packet, the result of multiplexing the first packet is used as the second packet, and the second packet is used as the second packet.
The packet assembling unit 40 of FIG. Therefore, in this case, what is output to the transmission line 2 or input from the transmission line 2 is the second packet, which operates similarly to the ATM network.
Hereinafter, the transmission line 2 will be described using an example of an ATM network.

Further, an ATM cell disassembling unit 50 for separating the multiplexed packet from the transmission line 2 into the line corresponding receiving unit 60 of the line corresponding packet receiving processing unit based on the channel identification information, and the existence of the separated packet. Extract the encoded voice information and silence detection information from the sound / silence identification information,
The information is sent to the voice decoding unit 62 of the means for decoding the encoded voice information or the silence reproducing unit 63 of the means for reproducing the silence from the silence detection information, and the decoded voice information or the silence reproduction information is transmitted. It has a depacketizing unit 61 of a unit for controlling a selecting unit 65 of a unit for selecting and outputting to the line from voiced / unvoiced information.

Further, in addition to means for realizing voice information communication between these devices, it has a function relating to processing when a packet is discarded during packet multiplexing. Specifically, the packet multiplexing control means 30 is provided. , A push button (PB) signal generator 64 for generating a packet discard notification signal to be output to the line side when it is determined to discard a packet on a certain line, and the already-selected decoded voice information or silence. It has a selection unit 65 which is means for selecting a packet discard notification signal from the reproduction information and outputting it to the line. The push button (PB) signal generator 64 outputs a warning sound or a caution sound if it is a sound that generates a noteworthy sound.
It may be another sound, for example a dual tone generator of a similar sound generator.

(2) Description of Operation In the first embodiment of the present invention, the voice information inputted from the line is discriminated between the voiced portion and the voiceless portion, and the voiced portion is always encoded. However, for the silent part, the silence detection information is generated only at the start of silence, and thereafter it is not encoded, but is packetized and then multiplexed and transferred, resulting in efficient voice transmission with a smaller transmission band. Can communicate.

Generally, the occurrence distribution of the silent state in a voice call has a statistical property, and therefore, when such communication is multiplexed, the transmission bandwidth varies. In the packet multiplex control process, when the amount of voice information packets exceeds the preset bandwidth of the transmission path, it is possible to prevent packets from being discarded due to a temporary increase in the transmission band. In order to delay the multiplexing process of a new conversation.

Specifically, when the number of packets to be multiplexed exceeds a predetermined upper limit value, the beginning and the beginning of a voiced packet newly generated from a silent state is accumulated in a buffer, and the packet transmission from the buffer is multiplexed. It starts after there is a vacancy in the number of packets being converted. This conversation is multiplexed via the buffer until the next silence occurs, so that there is always a delay during that period. Similarly, when a new conversation occurs, packet accumulation is also started in the buffer. However, if the buffer has insufficient free storage capacity, the packet is discarded and the packet discard notification signal is sent to the receiving side of the line. Is sent. Upon receiving the packet discard notification signal, the caller can pause the conversation and restart the conversation if the buffer storage capacity becomes available.

[0030]

[First Embodiment] (1) Description of Configuration In order to describe the embodiment of the present invention described above in more detail, a first embodiment of the present invention will be described in further detail with reference to the drawings. Explained.

FIG. 1 is a block diagram showing the configuration of a voice packet multiplex transmission apparatus according to an embodiment of the present invention. The voice packet multiplex transmission apparatus 1 shown in FIG. 1 performs voice code input and line compression on voice information input from a line, packet-multiplexes it, sends it out to a transmission line 2 as an ATM cell, and packet-multiplexes it from a received ATM cell. It is a voice packet multiplex transmission device that reproduces voice information after separation and sends it to a line.

The voice information input from the line is sent to the voice / non-voice detecting unit 13 of the line-corresponding transmitting unit 10 by
A voiced state and a voiceless state are discriminated for each predetermined section, and in the case of the voiced state by the voice encoding unit 11 according to the determination result of the voiced / non-voiced state, a code that is voice code compressed by PCM encoding or the like. In the case of a silence state, the silence compression processing generates silence detection information only at the start of silence, and the subsequent encoding is not performed. In the packetizer 12, channel ID information for identifying a channel and voice / non-voice information for identifying voice / non-voice are added to the encoded voice signal compressed by the voice code and the voice / non-voice detection information. , And voiced packets and silence detection packets, respectively.

The packet multiplexing control unit 30 monitors the number of channels to be multiplexed, and when the packet of the channel for which the multiplexing process needs to be delayed is accumulated in the packet buffer unit 20 and a new multiplexing process becomes possible. To start the packet transmission from the packet buffer unit 20 on the untransmitted channel.

The ATM cell assembling unit 40 multiplexes the packet from the packetizing unit 12 or the packet buffer unit 20 of each line according to the instruction of the packet multiplexing control unit 30, and further divides this into a plurality of ATM cells for transmission. The process of transmitting to the path 2 is performed.

The ATM cell disassembling unit 50, contrary to the ATM cell assembling unit 40 processed / transmitted by the executed line-corresponding transmission unit 10, is processed / transmitted in the same manner as described above, and an ATM cell of this kind is processed. A multiplexed packet is reconstructed from a plurality of received ATM cells, and the packet is output to the depacket unit 61 of each line-corresponding reception unit 60 based on the channel ID information added to each packet. The depacket unit 61 uses the voice / non-voice information added to the packet to the voice decoding unit 62 for voice packets and the voice reproducing unit 6 for voice detection packets.
3 based on the received packet by controlling the selection unit 65 which outputs the audio information compressed and decoded by the audio decoding unit 62 and the silence information reproduced by the silence reproducing unit 63. Outputs playback audio information to the line. In addition, the PB connected to the selection unit 65
The signal generation unit 64 is used for notifying the receiving side of the channel that the packet is discarded when the packet is discarded due to lack of free space in the packet buffer unit 20 in the packet multiplexing process by the packet multiplexing control unit 30. Controlled by the packet multiplexing control unit 30 together with the selection unit 65, a PB signal is output to the line when a packet discard occurs.

(2) Description of Operation Next, the operation of the voice packet multiplex transmission apparatus of one embodiment of the present invention will be described in more detail with reference to the drawings.

FIG. 2 is a flow chart for explaining the operation of the voice packet multiplex transmission apparatus of one embodiment of the present invention.

FIG. 3 is a diagram for explaining the packet multiplex processing of the voice packet multiplex transmission apparatus in one embodiment of the present invention.

When the voice information input from the line is in the voiced state, the line-corresponding transmitter 10 periodically outputs the voiced packets of the fixed length information block. In the silent state, a silence detection packet added with silence information of the fixed-length information block is output at the start of silence, but no packet is output thereafter.

With respect to the band limitation provided in the transmission line 2,
Since the information amount per unit cycle of the voiced packet or the silence detection packet output thereto is a fixed length as described above, the packet band limiting process in the packet multiplexing control unit 30 is performed by the cycle of the voiced packet output. Is set as a unit cycle, and the number of channels of packets to be multiplexed per unit cycle is controlled by monitoring the number of channels to be packet-multiplexed every period so that the number of channels of packets is equal to or less than a value obtained from the limited band. Here, the processing flow of the packet multiplexing control unit 30 will be described with reference to FIG.

The packet multiplex control unit 30 receives the voice packet or the silence detection packet output from the packetization unit 12 or the packet buffer unit 20 of each line in the process of determining the multiplexed channel in each cycle as shown in FIG. The multiplexed channel is determined according to the processing flow. First, in the first condition judgment (S11), regarding the channels already stored in the packet multiplexing control unit 30,
The latest packet is always stored at the end of the same channel in the buffer (S16) to guarantee that the packet output order does not change during multiplexing. However, when the free space of the packet buffer unit 20 is insufficient (S13), a PB signal indicating packet discard is sent to the receiving side of the channel (S17) and the packet is discarded (S1).
8).

In the second condition judgment (S12), when a voiced packet is generated in a new channel, the band increases. Therefore, the number of channels to be multiplexed in this cycle is verified to continue from the previous cycle. The total of the number of channels to be sent out and the number of unsent channels held in the packet buffer unit 20, that is, the total of the head voice packet, the number of continuously sent channels, and the number of buffer reserved channels is larger than the maximum number of multiplexed channels. Only when the number of channels is less than the maximum number of channels that can be tolerated (No), multiplexing is started for the number of vacant channels, and the packet buffer section 20 is used for more channels. accumulate.

On the other hand, if the sum of the head voiced packet, the number of continuous transmission channels and the number of buffer hold channels is larger than the maximum number of multiplexed channels (Yes), then it is determined whether or not there is a vacancy in the packet buffer unit 20. If the free space of the packet buffer 20 is insufficient (S14) (No), a PB signal indicating packet discard is sent to the receiving side of the channel (S17), and the packet is discarded (S18). . If there is free space in the packet buffer unit 20 in step S14, the packet buffer unit 20
The packet is stored in the area of the new channel in the (S1
5).

By the second condition judgment, it is ensured that the channels continuing from the previous cycle are preferentially multiplexed, and the channels generated later are packet buffer unit 20.
This prevents the priority of the channel that has been suspended for sending.

In the third condition judgment (S19), if the number of channels that are continuously transmitted is less than the maximum number of multiplexed channels (Yes), untransmitted channels that are held in the packet buffer unit 20. Multiplexing is started (S20). If the number of channels to be continuously transmitted exceeds the maximum number of multiplexed channels (No) in the third condition determination (S19), the ATM cell assembling unit 40 performs step S
With the result of 20, the packet is multiplexed and sent to the transmission line 2 as an ATM cell (S21).

An example of this multiplexed channel determination processing will be described with reference to FIG. FIG. 3 shows an example in which the maximum number of multiplexed channels is set to 3 in an apparatus having four channels of inputs (CH1 to CH4). Also, the voiced packets (each 1 to 4 CH) and silence detection packets (each 1 to 4 CH) of each channel and the silence state are distinguished as shown in the figure.

In this example, a certain period (T1 to T1
5. ． ) Is multiplexed in the next cycle and output as an output multiplex voice packet. Channel CH1 generated in the T6 cycle to the T10 cycle
Since the voice packet or the silence detection packet of No. 1 exceeds the maximum number of multiplexed channels at the time of multiplexing, it is temporarily stored in the packet buffer unit 20 as shown as a reserved packet in the packet buffer unit 20 of FIG. It is output with a delay of two cycles. However, due to the influence of the channel CH1, packet discard, packet fluctuation, etc. do not occur in the other channels.

Here, an example of the operation by the packet multiplexing means, which is often implemented in the conventional voice packet transmission device,
This will be described with reference to FIG. This drawing shows an example of an output packet when two different multiplexing methods are used under the same input condition and the limitation of the maximum number of multiplexed channels as in FIG. Figure 6
The first multiplexing example in FIG.
Repeated multiplexing in the order of H4 except for silent packets,
When the cycle changes, it is a method of multiplexing from the next channel multiplexed in the previous cycle, and packets exceeding the maximum number of multiplexed channels are discarded. In this example, when the maximum number of multiplexed channels is exceeded, packets are dropped in the channels being communicated with almost the same probability, which causes the communication quality to be degraded almost equally in all channels. Become.

On the other hand, in multiplexing example 2 shown in FIG. 6, the order of multiplexing is the same as that in multiplexing example 1, but packets exceeding the maximum number of multiplexed channels are accumulated in the queuing buffer,
This is a method of preferentially transmitting this packet in the next cycle. In this example, when the maximum number of multiplexed channels is exceeded (T7 to T11), delay fluctuations of packets occur at almost the same probability in the channels being communicated, and the receiving side device absorbs the fluctuations. If not, the quality of communication with the device will deteriorate, or if the maximum number of multiplexed channels is exceeded and packet drops continue to occur, all channels will be evenly distributed. Communication quality will be degraded.

In contrast to the multiplexing examples 1 and 2 shown in FIG. 6, in the present embodiment, as shown in FIG. 3, in order to keep the communication quality above a certain level, each block configuration explained in FIG. With the functions of the flowchart described with reference to FIG. 2, it is possible to prevent the discarding of packets, the fluctuation of packets, and the like in the channel due to the mutual influence of the channels as described above.

[Second Embodiment] As a second embodiment of the present invention, an ATM (Asynchronous Transfer Mod) shown in FIG.
e: Asynchronous communication mode) The cell assembling unit 40 and the ATM cell disassembling unit 50 are connected to the transmission line 2 through the ATM.
Although it is shown as a functional unit for transferring a voice packet using a cell, the S that has been generally realized conventionally.
When a voice packet is transmitted using a TM (Synchronous Transfer Mode) -based packet transfer means, an AT is used.
The M cell assembling unit 40, as the STM packet assembling unit 40 ′, only performs multiplexing of voice packets, and similarly
The M-cell disassembling unit 50, as the STM packet separating unit 50 ′, distributes the voice packets from the multiple voice packets to the line reception corresponding unit 60 based on the channel ID information. Also, the PB signal generation unit 64 also generates a busy tone BT signal that has been transmitted for a disconnected call at the time of call setting, in a conventional STM-based switching system or the like, as the BT signal generation unit 64 '.

Here, the STM method is the world-uniform BI.
SDH (Synchronous Digital Hierarchy) that unifies the digital hierarchy to establish the SDN standard,
The interface between the multiplexer and the transmission line is NNI (Ne
PDH before SDH as twork Node Interface)
(Plesiochronous Digital Hierarchy) is also included, and the structure of a transmission signal in which various signals are multiplexed is a synchronous transmission module (STM: Synchronous Transport Module).
Based on, the multiplicity N is used and STM-N (N = 1,
4, 16), the STM-N has a structure of 270 × N bytes × 9 rows, which is accommodated in 125 μs, and the information area is transmitted as an STM-N payload.

Therefore, the packet assembling unit 4 according to the present embodiment.
The multiplexed STM frame at 0 ′ is assembled in a predetermined cycle in 125 μs units. Also, instead of the ATM cell disassembling unit 50, a packet disassembling unit 50 ′ disassembles a desired packet and sends it to the depacketizing unit 61.

[Third Embodiment] Further, as a third embodiment of the present invention, since the voice packet is transferred to the transmission line 2 in FIG. The IP packet assembling unit 40 ″ performs multiplexing of voice packets and addition of IP header information, and the ATM cell disassembling unit 50 acts as an IP packet separating unit 50 ″ on the Internet for separating multiple voice packets from IP packets and channel ID information. The voice packet is distributed to the line reception corresponding unit 60 based on the above.

It should be noted that the Internet handles multimedia that transmits not only voice but also image signals and data, and TCP / IP (Transmission Control Protocol / In
ternetProtocol) protocol, fixed-length or variable-length data structure may be transmitted, and in the case of voice or image, real-time transmission may be required. Internet packets according to the digital data structure are IP packets. It is preferable that the assembling unit 40 ″ and the IP packet separating unit 50 ″ deal appropriately.

In this case, for example, the packetizing unit 12
The packet handled by the depacketizing unit 61 is treated as the first packet, and the packet transmitted or input by the IP packet assembling unit 40 ″ and the IP packet separating unit 50 ″ is treated as the second packet.

[0057]

As is apparent from the above description, the present invention has the following effects.

The first effect is that when the amount of voice packets to be multiplexed and transmitted to the transmission path exceeds a predetermined band capacity, the timing at which multiplexing is started for a channel in which a new voice packet is generated. In order to delay the delay, packets are temporarily stored in the packet buffer, and after the bandwidth capacity becomes available, multiplexing is started, so that the overall communication quality and voice quality are not significantly affected, and temporary This has the effect of preventing packet discard as much as possible by increasing the number of multiplex channels.

That is, since a voice packet of a certain channel is generated, delay fluctuation in the waiting for multiplexing and packet discarding do not occur in the channel that has been multiplexed until then, and the packet is temporarily stored in the packet buffer. This is because the packets of the channel to be output are also transmitted at regular intervals without delay fluctuations waiting for multiplexing after the start of multiplexing.

The second effect is that it is possible to reduce the buffer capacity prepared for absorbing the packet delay fluctuation in the receiving side device. This is because the receiving side device converts packets whose arrival intervals fluctuate to some extent into voice signals, so a buffer is required to temporarily store the incoming packets and absorb the fluctuations in packet delay. The smaller the number, the smaller the buffer capacity for accumulating packets.
The reason for this effect is a packet multiplexing control method in which packets that cannot be multiplexed when voice packets exceed the bandwidth capacity are stored in a buffer, and the selection of channels to be stored in the buffer is switched regardless of changes in the silent / voiced state. In the case of, the delay caused by buffer accumulation occurs by switching to accumulation in the buffer at the time of multiplexing, but occurs as packet delay fluctuation after multiplexing.However, with the method of packet multiplexing control in the present invention, during voice packet communication Since this switching does not occur, packet delay fluctuation due to multiplexing is reduced.

A third effect is that the sender can communicate because a means for notifying the sender of the packet discard is provided when the packet is discarded due to exceeding the storage capacity of the packet buffer. It is possible to grasp a conversation that has not been performed and reduce voice packets that are retained in the packet buffer by interrupting the conversation, so that the packet buffer can be efficiently used.

That is, in the device having no packet discard notification means according to the present invention, the occurrence of packet discard is recognized by the communication partner as deterioration of voice quality or interruption of conversation, and the notification from the communication partner is given. Except for the fact that the sender cannot directly recognize the state of communication, the state of transmission of transmitted information could not be sufficiently grasped, especially when packet discard frequently occurred. , Because these are improved. In addition, the sender that received the packet discard notification suspends the conversation and stops the accumulation of further voice packets in the packet buffer, which reduces the traffic on the transmission path and accumulates in the packet buffer. This is because the existing packet was sent out, so that it becomes possible to newly store voice packets of another channel in the empty packet buffer.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of a voice packet multiplex transmission device of the present invention.

FIG. 2 is a flowchart for explaining the operation of the voice packet multiplex transmission apparatus in one embodiment of the present invention.

FIG. 3 is a diagram for explaining packet multiplex processing of the voice packet multiplex transmission apparatus in one embodiment of the present invention.

FIG. 4 is a block diagram showing a configuration of a conventional voice packet transmission device.

FIG. 5 is a block diagram showing a configuration of a conventional voice packet transmission device.

FIG. 6 is a diagram for explaining an example of the operation of the packet multiplexing means often implemented in the conventional voice packet transmission device.

[Explanation of symbols]

1 Voice packet multiplex transmission device 2 transmission lines 10-line compatible transmitter 11 Speech coding unit 12 Packetization unit 13 Sound / silence detector 20 packet buffer 30 packet multiplexing controller 40 ATM cell assembly department 50 ATM cell disassembly unit 60 line compatible receiver 61 Depacketizer 62 speech decoding unit 63 Silent playback section 64 PB signal generator 65 Selector 201 bit string conversion circuit 202-bit position number addition circuit 203 Packet information generation circuit 204 Packetizing circuit 205 silence detection circuit 206 control circuit 210 line support 211 Input terminal 221 Input audio signal 223 bit string conversion control signal 224 bit position information 230 Multiplexer

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Naoaki Tadokoro, 1-4-2 Mita, Minato-ku, Tokyo Inside Nippon Electric Telecommunication Systems Co., Ltd. (56) Reference JP-A-5-244201 (JP, A) Kai 2000-165446 (JP, A) JP Hei 8-340337 (JP, A) Murakami, K., Minowa, M., Futami, Y., E., Y., Toyama, T., Low Bit Rate Voice Packet Multiplexing, 1999 Annual University of Science, March 8, 1999, B-6-38 (58) Fields investigated (Int.Cl. ⁷ , DB name) H04L 12/56 H04L 12/66 H04M 3/00 H04M 11/00

Claims (4)

(57) [Claims] 1. A voice packet multiplex transmission apparatus that multiplexes and transmits voice packets via a line, comprising: a voice / non-voice detection unit for determining a voiced portion and a voiceless portion in voice information input from the line. , The voiced part is encoded based on the discrimination result of the voice / silence detection part, the silence detection information is generated only at the start of silence for the silent part, and is not encoded in the silent state thereafter. a speech coder, the adds the information identifying the information and voice or silence identifies the channel to the coded audio information is output from the speech encoding unit and the silence detection information Packet 1
A packetizing unit for packetizing the packet into a packet, a packet buffer unit for accumulating the first packet to delay the start of multiplexing of the first packet output from the packetizing unit, and the packetizing unit or the second to output the first packet multiplexing turned into a second packet of the packet to form a second the output from the packet buffer to the transmission path
Packet assembling unit and the first packet that is multiplexed by the second packet assembling unit.
Monitor the number of channels corresponding to
Channels to delay the start of multiplexing based on the number of channels
The first packet of Le performs control of storing in the packet buffer unit, also accumulated is in the packet buffer unit
Is a packet multiplexing controller for controlling the multiplexing start of the first packet is, said towards the sender of the first packet first packet
Push button (PB) signal to notify the discard of
PB signal generator that outputs the signal to the line corresponding to the sender
And the packet multiplexing control unit sends the first packet to the
When storing in the packet buffer, the packet buffer
If there is no free space in the header part, the first packet is discarded.
Discard the packet and send the packet corresponding to the sender of the first packet.
Instruct the PB signal generator to output the PB signal to the line
A voice packet multiplex transmission device characterized by: 2. A voice coding unit, a voice / non-voice detecting unit, and a voice detecting unit.
Line-corresponding transmitter including a packetizing unit, and the PB
In the second packet reception processing means including a signal generator,
A line-corresponding receiver is provided for each line, and further, the ATM packet which is the second packet is transmitted from the transmission line.
The ATM session that receives the packet and decomposes and outputs it to the receiving unit corresponding to each line.
The ATM cell disassembling unit,
A depacketizing unit for depacketizing voice packets, and
Before receiving the audio digital signal from the depacketizer
Silence playback section that detects silence information and plays silence information
The voiced sound from the depacketizer
Voice decoding unit for outputting information, and the decoded voice
Abolition of the information, the reproduced silence information and the first packet
The PB signal for notifying the abandonment is sent from the packet multiplexing controller.
And a selection unit for selecting and outputting to the line by the control of
Voice packet multiplexing transmission apparatus according to claim 1, characterized in that. 3. The packet multiplexing controller is configured to control the packet
Same as the channel of the first packet to be stored in the buffer
The first packet of one channel is already in the packet buffer
The first package to be stored when being stored in the
Already stored in the packet buffer section.
The voice packet multiplex transmission apparatus according to claim 1 or 2 , wherein the first packet of the same channel is stored at the end of the first packet . 4. A voice packet is multiplexed and transmitted via a line.
In the voice packet multiplex transmission method, the voice information input from the line is divided into a voice part and a voice part.
The sound part is discriminated, and the voiced part is encoded based on the discrimination result.
Go and output it as voice information.
It encodes only when silence starts and outputs silence detection information.
In the silent state after the above, no coding is performed, and the line is connected to the voice information and the silence detection information.
Information that identifies the indicated channel and information that identifies voiced
And packetizing into a first packet, and starting the multiplexing of the packetized first packet.
Accumulating the first packet for delaying, the packetized first packet or the accumulating
The first packet that has been multiplexed is converted into an ATM cell.
The number of channels corresponding to the multiplexed first packet
And multiplex based on the number of channels in the monitoring
Store the first packet of the channel that should delay the start of
The first package that controls the stacking and is accumulated.
When the first packet is stored by controlling the start of packet multiplexing, the free space used for the storage
If there is no capacity, discard the first packet.
The PB signal to the line corresponding to the sender of the first packet.
A voice packet multiplex transmission method characterized by outputting a signal.